Moon,
No, I do not see anything like that. Everything in zeppelin.log on all nodes
looks fine.
My log4j.configuration on my “main” node (Zeppelin) is
log4j.rootLogger = INFO,console,dailyfile
log4j.appender.console=org.apache.log4j.ConsoleAppender
log4j.appender.console.target=System.err
log4j.appender.console.layout=org.apache.log4j.PatternLayout
log4j.appender.console.layout.ConversionPattern=%d{yy/MM/dd HH:mm:ss} %p %c{1}:
%m%n
log4j.appender.dailyfile.Threshold = DEBUG
log4j.appender.dailyfile = org.apache.log4j.DailyRollingFileAppender
log4j.appender.dailyfile.File = /opt/zeppelin/logs/zeppelin.log
log4j.appender.dailyfile.layout = org.apache.log4j.PatternLayout
log4j.appender.dailyfile.layout.ConversionPattern=%5p [%d] ({%t} %F[%M]:%L) -
%m%n
By default dailyfile.File is a property, which I’ve just changed manually.
My understanding was that the path it tries to write on all nodes of my
cluster. And that’s the reason why everything became fine as soon as I copied
Zeppelin on all worker nodes. Without the copy, it just couldn’t create that
directory (it’s in /opt/ and Worker doesn’t have sudo permissions).
On 21 May 2015, at 03:28, moon soo Lee <[email protected]> wrote:
> Hi,
>
> Zeppelin suppose not to be installed in all modes of Spark cluster.
> Do you see any exception in Zeppelin's log file?
>
> Thanks,
> moon
> On 2015년 5월 20일 (수) at 오후 10:26 Eugene Morozov <[email protected]> wrote:
> Well,
>
> it looks like Zeppelin must be installed onto all nodes of Spark cluster.
>
> It might happen that it’s just required to create logs directory for zeppelin
> so that all worker nodes are able to put logs into configured file, but it
> also might happen that the whole Zeppelin is required.
>
> Could you, please, confirm?
>
> On 20 May 2015, at 12:57, Eugene Morozov <[email protected]> wrote:
>
>> Hi!
>>
>> I can’t understand where I can get logs from my workers (failed executors)
>> when I’m running job from Zeppelin.
>> I’m using the most recent version of Zeppelin.
>>
>> In spark worker stderr I see the following
>>
>> 15/05/20 09:23:36 INFO Executor: Fetching
>> http://x.x.x.x:51149/jars/my-jar.jar with timestamp 1432113678510
>> 15/05/20 09:23:36 ERROR CoarseGrainedExecutorBackend: Driver Disassociated
>> [akka.tcp://[email protected]:49252] ->
>> [akka.tcp://[email protected]:33718] disassociated! Shutting down.
>>
>> In zeppelin.log itself:
>> Lost task 5.0 in stage 0.0 (TID 5, x.x.x.y): ExecutorLostFailure (executor 0
>> lost)
>>
>>
>> But where to get the Exception stack trace of the Executor? They seems to be
>> not appear in stdout nor stderr. Stdout is completely empty and stderr does
>> not contain anything helpful.
>>
>> Please, advice.
>> --
>> Eugene Morozov
>> [email protected]
>>
>>
>>
>>
>
> Eugene Morozov
> [email protected]
>
>
>
>
Eugene Morozov
[email protected]
